Using Emotional Signals to Make Sense of People's Actions – Autism and Typical Development

Background: From infancy, typically developing children understand and predict people’s actions relying on emotional signals. For example, they expect an agent to pick up an object that they are looking at with a happy or satisfied expression, rather than an object that they are looking at with a disgusted expression. Literature indicates that children with autism show abnormalities in both attending to people’s faces and interpreting emotional signals. We investigated to what extent children with autism are sensitive to emotional signals and use them to predict an agent’s behavior.

Objectives: We tested two alternative hypotheses:

(1)               Children with autism will fail to predict an agent’s behaviour as a consequence of diminished visual attention to changes in the agent’s emotional expressions

(2)               Children with autism will fail to predict an agent’s behaviour as a consequence of difficulties in interpreting such emotional signals

Methods: 18 children with autism and 18 typically developing subjects matched for IQ and age observed a series of videos showing an actor performing actions on objects. The videos stopped before the action was completed and participants were asked to complete the observed action. In the experimental condition the actor’s behavior could be predicted only by considering her emotional expressions. For example, in one trial the actor was choosing what objects to put in a container and while doing so she was looking at certain objects with a happy expression and to other objects with a disgusted expression. In the control condition the actor’s emotional expression was neutral and her action could be predicted based on the characteristics of the objects. During the observation of the videos, participants’ eye movements were recorded using an eye-tracking system.
Results: Both groups predicted the agent’s behavior based on the objects’ characteristics in the control conditions. Contrary to our hypotheses, both groups predicted the agent’s behavior based on the agent’s emotional expressions in the experimental condition. Children with autism looked at the agent’s face less frequently than participants in the control group in the neutral condition. However, they looked at the agent’s face as much as controls in the experimental condition.

Conclusions: These data suggest that emotional cues trigger attention to the agent’s face in children with autism, and are used successfully by them to predict an agent’s behaviour. Implications for treatment will be discussed.